Cardiovascular Disease Risk Calculator

A web-based clinical decision support tool that uses a Bayesian Network to predict the risk of cardiovascular disease. The application provides real-time risk assessment based on patient attributes and visualizes both the network structure and the influence of various factors.

Features

• Real-Time Risk Calculation: Instantly calculates and displays the probability of heart disease as patient attributes are updated.
• Bayesian Network Model: Utilizes a 5-tier Bayesian Network trained on the UCI Heart Disease dataset to model probabilistic relationships between clinical factors.
• Interactive Visualization: Dynamically renders the Bayesian Network, coloring nodes based on their calculated probabilities to provide an intuitive understanding of the model's state.
• Top Risk Factors: Identifies and highlights the top three most influential factors contributing to the current risk assessment.
• Detailed Insights: Provides probabilities for each individual factor, showing how it contributes to the final risk score.
• Responsive UI: Built with React, Tailwind CSS, and shadcn/ui for a modern and responsive user experience.

How It Works

The application's core is a Bayesian Network model whose structure and Conditional Probability Tables (CPTs) were derived from a Jupyter Notebook using the pgmpy library. This model logic has been ported to a TypeScript-based Supabase Edge Function for scalable, serverless execution.

1. Data Input: The user inputs 13 different patient attributes (e.g., Age, Blood Pressure, Cholesterol) through a series of dropdowns on the web interface.
2. API Call: Each change in the input data triggers an API call to a Supabase Edge Function, sending the complete patient profile.
3. Backend Inference: The Supabase function takes the patient data and applies it to the pre-computed Conditional Probability Tables (CPTs). It calculates P(Disease | Factor) for 10 parent factors.
4. Risk Aggregation: A weighted average of these individual probabilities is computed to produce a final, aggregated risk score. This method serves as a fast approximation of full Bayesian inference.
5. Result Generation: The function returns a JSON object containing the final disease probability, the probabilities contributed by each factor, and the top three most influential factors.
6. UI Update: The React frontend receives the results and dynamically updates the UI components:
   • The Risk Dashboard shows the main probability percentage.
   • The Top Factors panel lists the key contributors.
   • The Network Visualization colors each node based on its probability and adds a glowing border to the most influential nodes.

Technology Stack

• Frontend: React, Vite, TypeScript, Tailwind CSS, shadcn/ui
• Backend: Supabase Edge Functions (Deno/TypeScript)
• Data Visualization: Recharts, SVG
• Data Analysis & Modeling: Python, Jupyter, Pandas, pgmpy

Data Source

The model was trained on the Heart Disease Data Set from the UCI Machine Learning Repository. Initial data exploration and preprocessing were performed in the SourceCode/Jorjit/jupyter/eda.ipynb notebook. The dataset was binned into categorical values as detailed in SourceCode/Jorjit/jupyter/Data_Binning.ipynb.

Getting Started

To run this project locally, you will need Node.js and npm installed.

1. Clone the Repository

git clone https://github.com/jorjitdasoria/cardiovascular-disease-calculator.git
cd cardiovascular-disease-calculator

2. Navigate to the Application Directory

The main application code is located in the Interim Code folder.

cd "SourceCode/Interim Code/cardiovascular-risk-calculator-main/cardiovascular-risk-calculator-main/"

3. Set Up Environment Variables

Create a .env file in the root of the application directory and add your Supabase project credentials. You can use the existing values if you do not have your own Supabase project.

VITE_SUPABASE_PROJECT_ID="vywbtotrjwbdfgiqosnb"
VITE_SUPABASE_PUBLISHABLE_KEY="eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJzdXBhYmFzZSIsInJlZiI6InZ5d2J0b3RyandiZGZnaXFvc25iIiwicm9sZSI6ImFub24iLCJpYXQiOjE3Njc2MTA3NzEsImV4cCI6MjA4MzE4Njc3MX0.Es7SkCsjIGT_YKu5RGUONfqkfLBVj7-Ds_RCrRs2s8A"
VITE_SUPABASE_URL="https://vywbtotrjwbdfgiqosnb.supabase.co"

4. Install Dependencies

npm install

5. Run the Development Server

This will start the Vite development server, typically on http://localhost:8080.

npm run dev

6. Deploying the Supabase Function

The backend logic resides in a Supabase Edge Function. To deploy it, you will need the Supabase CLI.

# Link your local repository to your Supabase project (if you created one)
# supabase link --project-ref <your-project-id>

# Deploy the function
supabase functions deploy calculate-risk --no-verify-jwt